I LL I N
~
S
UNIVERSITY OF ILLINOIS AT URBANA-CHAMPAIGN
PRODUCTION NOTE University of Illinois at Urbana-Champaign Library Large-scale Digitization Project, 2007.
317?. 1•-
T
CE I p C O AR L T S
Technical Report No.
66
The Development of Strategies for Studying Prose Passages" Ann Lo Brown and Sandra S.
Smiley
University of Illinois at Urban~a-Champaign October 1977
Center for the Study of Reading S• RAORY FTN 'A. OCT 719 7 )1t
up
UNIVERSITY OF ILLINOIS AT URBANA-CHAMPAIGN 1005 West Nevada Street Urbana, Illinois 61801
BOLT BERANEK AND NEWMAN INC. 50 Moulton Street Cambridge, Massachusetts 02138
~I~
II
t~L~CLC~'
~-IILICL~
The person charging this material is responsible for its return to the which it was withdrawn on orlibrary from before the Latest Date stamped below. Theft, mutilation, and underlining of books are reasons for disciplinary action and may result in dismissal from the University. To renew call Telephone Center, 333-8400 UNIVERSITY OF ILLINOIS rIRDoAV ....
i
CENTER FOR THE STUDY OF READING
66
Technical Report No.
The Development of Strategies for Studying Prose Passages
1
Ann L. Brown and Sandra S. Smiley University of Illinois
at Urbana-Champaign
October 1977
University of Illinois at Urbana-Champai gn 51 Gerty Drive 61820 Champaign, Illinois
Bolt Beranek and Newman Inc. 50 Moulton Street 02138 Cambridge, Massachusetts
This research was supported in part by Grants HD 06864, HD 05951, and a Research Career Development Award HD 00111, from the National Institutes of Child Health and Human Development and in part by the National Institute of Education under Contract No, MS-NIE-C-400-76-0.16.
Abstract
The strategies of children and college students were examined as they attempted to study texts. instructions,
College students,
under various intentional learning
displayed a repetitive diagnostic pattern.
Following extended
study they improved recall of important, but not unimportant, Eleventh and twelfth graders conformed to the adult pattern, eighth graders were not as efficient.
elements of texts. but fifth through
Older students benefitted from increased
study time because they possessed the necessary knowledge concerning the importance of text segments to enable them to concentrate on the essential.
Younger
students, not so prescient, do not concentrate exclusively on the important units, for they did not know what they were. Age was not the sole determinant of performance for some students at each age spontaneously adopted the strategies of underlining or note-taking.
Those
who did, concentrated on the important elements and subsequently approached the adult-like pattern in even if
recall; those who did not,
induced to adopt one of the strategies.
displayed the immature pattern, The interplay of knowledge
concerning texts, study strategies and effective recall was described.
2 Getting the gist of a message, whether oral or written, is communicative activity.
Without this ability,
an essential
children would never learn a
language and would certainly never come to use that language as a vehicle for communication.
Extracting the main idea is
clearly an essential information-
gathering activity and the ability to glean the main message,
to the exclusion
of nonessential detail, must be a naturally occurring ability, given of course, a reasonable match between the complexity of the message and the receiver's current cognitive status (Brown,
1975,
1978b; Brown & DeLoache,
In a series of recent studies (Brown & Smiley, Townsend,
& Lawton,
1977; Smiley, Oakley,
Worthen,
19 7
7
1978).
; Brown,
Campione,
Smiley,
& Brown,
Day,
1977),
we have been concerned with children's ability to extract the main theme of prose passages.
Our interest in
and practical grounds.
First,
mature information processor is
this topic can be defended on both theoretical
there is
adept at channelling his attention to the most
informative aspects of the stimulus. find it is
considerable evidence that the more
Conversely,
young children or novices
more difficult to ignore irrelevant or less-informative material.
true whether the task involves visual scanning (Brown & DeLoache,
Mackworth & Bruner, Zinchenko, 1973),
1970; Pushkina,
Chzhi-tsin & Tarakanov,
1971; Thomas, 1963),
1968: Vurpillot,
This
1978;
1968;
selective attention (Hale & Piper,
or incidental learning paradigms (Hagen,
1972; Postman,
subject must selectively ignore aspects of the stimuli.
1964) where the
An extension of these
findings to situations where children must study prose passages would be theoretically interesting. Perhaps of more importance is
the practical aspects of this work.
Much
of what we are required to learn must be extracted from prose passages; studying prose is
the leading activity (Brown,
1978a) of schools.
Effective
reading and studying both involve the ability to extract the essential message
3 and discard trivia, as indeed does effective listening.
Thus the current
interest in understanding and remembering prose can be seen as a direct result of the call for ecological validity raised by memory theorists, developmental or otherwise (Bransford, Franks, Morris & Stein, 1978; Brown, 1978a, 1978b). To date we have shown that even kindergarten children (Smiley et al, 1977), educable retarded grade schoolers (Brown & Campione, 1978), and poor readers (Smiley et al.,
1977) are sensitive to the importance of various sections of
texts, for their recall scores reflect the rated importance of the constituent idea units of stories.
Although older children (seventh grade) did recall more
than younger children (third grade) in the original Brown and Smiley (1977) study, there was no interaction of age and importance level.
Children at each
age tested recalled more important than unimportant elements of the text. In the Brown and Smiley study the students were also asked to rate the importance of textual elements to the theme of the entire story.
Here a dramatic
age effect was found, for there was a gradual improvement in the ability to identify the important sections.
College students could separate units of text
into each of the four levels of importance, previously identified by other groups of college students, a nice replication factor.
Seventh graders could
separate low, medium, or high levels but were insensitive to fine gradations at the medium levels of importance.
Fifth graders were only able to isolate
the most important units from the remaining three levels, while third graders made no reliable distinction between levels of importance. Younger children's ratings cbuld diverge from those agreed upon by adults either because the children judged different material to be important or because they were not consistent in their importance rating patterns. latter appears to be true.
The
The ratings of the younger children were idio-
syncratic with most units receiving the full range of possible scores.
As
4 children were not asked to rate and recall the same passage, it was not possible to consider whether an individual child's rated importance was related to his own recall selectivity.
But, in view of the close correspon-
dence between the recall patterns of subjects of all ages, it was assumed that the rating patterns shown by the younger children reflected their insensitivity to degree of centrality to the theme rather than a different (and consistent) impression of what material was important. A further reason why the young children did not display sensitivity to importance level could be the particular rating task chosen.
Students were
required to read (or listen to) the whole text and then eliminate (cross out) one quarter of the idea units.
This procedure was repeated twice until only
one quarter, the most important, remained.
This is a reasonably difficult
task, and even though the younger children received considerable pretraining, the difficulty of the task may have obscurred their sensitivity to fine degrees of importance. through twelfth grade,
To check this hypothesis, we asked children from fifth and college students to read the same stories and then
to pick thel2most important units or to pick the 12 units they would like for retrieval cues (Brown,
Smiley & Lawton, 1977).
Children from seventh
grade and above chose almost all level 4 units (the most important) and few level 1 units: 88% of college choices and 73% of seventh grade choices were of level 4 units.
Fifth graders, however, selected only 48% of the level 4
units and their remaining choices were randomly distributed across the other three levels of importance.
Thus, even with the easier task, fifth graders
were only able to differentiate the most important levels from all others. As children mature, they become better able to identify the essential organizing features and crucial elements of texts.
Yet this must be an
essential prerequisite for effective use of a limited processing capacity and
5 limited time when studying.
Without such knowledge, it would be difficult for
the child to select important units for extra processing.
The adult reader,
however, thanks to his foreknowledge concerning the relative importance of sections of the material he is studying, should be able to make effective use of extended study time.
In order to concentrate on the essential at the
expense of trivia, one must know what the essential elements are. We tested this hypothesis in the following experiments.
In the first
section, we used only college students, and considered their performance under various intentional learning and incidental orienting conditions.
If it is
true that knowledge concerning the importance of certain textual elements leads to effective study, then we would predict a specific diagnostic pattern of recall scores.
Following intentional study, the college student should
improve his recall, but this improvement should be differentially distributed across the various degrees of importance.
The effective learner should not
direct extra study to the trivial units and, therefore, one would not expect an increase in recall of nonessential information.
Concentrating the focus of
his efforts on the important elements of the story, the efficient studier should enhance his recall of essential material. The final experiment is a developmental study. twelfth grade were also allowed extra study time.
Students from fifth through The relationship between
their knowledge of textual importance and their knowledge of effective study strategies was examined with reference to the diagnostic pattern of recall scores and the physical records they-produced, underlining or note-taking, while studying.
The main prediction is that there should be an intimate relationship
between the subject's knowledge of the importance of specific units of texts, his knowledge of strategies, and his ability to benefit from additional study time.
If young children are not aware of the degree of centrality of a text
6 unit to the theme of a story, they can scarcely be expected to select out important units for extra study. Experiment la Method Subjects.
The subjects were 80 college student volunteers paid $2.00
for their participation. Stimulus materials. reported here.
Half the subjects were female. The same stories were used in
They were two Japanese
"How to Fool a Cat",
2
folk tales,
all
of the experiments
"The Dragon's Tears" and
selected because of their formal similarity; they both
featured a trick ending and could be described as conveying a moral.
Further-
more, both stories have been found to interest very young readers (Smiley et al.,
1977),
but they are still
(Brown & Smiley, 430 words,
1977).
This is
fifth
In addition, they are of comparable length (390 and
34 and 28 lines) and contain approximately the same number of
idea units (59 and 54). (i.e.,
suitable for presentation to an adult population
Finally,
they are of comparable readability levels
grade level, Dale-Chall readability scores of 5.2287 and 5.3682).
an important control for developmental studies,
for even the youngest
children studied here would be able to read them. The stories were divided into subunits following a procedure used by Johnson (1970)
and Brown and Smiley (1977).
Twenty-one college students were
asked to read the stories thoroughly and then to divide the text into individual units by placing a vertical
line
at a division point.
An individual unit was
defined as one that contained an idea and/or represented a pausal unit, a place where a reader might pause.
i.e.,
After division into independent units,
each story was retyped with one unit per line, and a second group of 34 college students was asked to rate the importance of each unit to the theme of the story using a four-point scale.
First they were asked to eliminate
7 one quarter of the units that they judged to be least important to the theme of the passage.
This procedure was then repeated twice more until only one
quarter of the units remained.
These last rmmaining units were judged the
most important to the theme, while the set eliminated first were the least important.
(For fuller details of the rating procedure, see Brown & Smiley,
1977.) Procedure.
The students were tested in small groups or individually,
depending on scheduling.
All subjects first listened to a tape recording of
one of the stories (stories counterbalanced across treatment groups), they simultaneously read a printed version. four treatment groups.
while
They were randomly assigned to
Half the students received an immediate test as soon
as the written version of the story was removed.
The remainder were permitted
five minutes interaction with the story prior to their recall attempt.
The
students were further subdivided into those receiving incidental and those receiving intentional instructions.
Prior to hearing a tape recording of the
story, the incidental group were told that we were collecting foreign folk tales that illustrated traditional morals (likeAesop's fables).
We intended to use
the stories to study moral development in children cross-culturally.
They
were to listen to the story and then we would ask them to answer a questionnaire concerning the moral of the story.
For the immediate group, as soon as the
story ended we asked them to recall the gist in
their own words:, the delayed
group was given the stories to consider and asked to write a brief commentary on the moral and the suitability of the story for children in third to seventh grade.
After five minutes of this activity they were given a surprise recall.
The intentional group received explicit instructions that they must attempt gist recall.
The immediate group were tested for recall after hearing the
story, with no chance to study;
the delayed group received five minutes extra
8 study with the written passage, in
and were told to do anything they wanted to do
order to improve recall. The written protocols were coded and then scored for gist recall by two
independent raters (interrater reliability = .94). or not the gist of each idea unit was retained,
The judges rated whether
irrespective of the wording.
Results and Discussion Preliminary inspection of the data revealed no differences attributable to sex of subject or to story, and therefore these variables were not entered into the analyses.
The mean proportion correct recall as a function of treatment
group and importance level are illustrated in Figure 1.
Apparently,
the
Insert Figure 1 about here
intentional group was better able to make use of the extended interaction with the story than were the incidental group. 2 (Immediate-Delay)
x 4 (Importance
A 2 (Intentional-Incidental)
Level) mixed analysis of variance revealed
significant main effects of Intentionality,
_F (1,76) = 6.02,
in the intentional group outperformed subjects in < .01.
F (1,76) = 7.23,9 F (3,228) = 309.35,9
x
R < .025.
Subjects
the immediate condition,
The main effect of importance level was also reliable,
< .001,
with recall an increasing function of importance
level. Of more interest, the following interactions were also reliable, Delay x Importance Level, Level, F (3,228)
= 3.29,
= 5.30,
F_ (3,228) R < .025;
j
< .005;
and the three-way
Delay x Intentionality x Importance Level was just reliability,
F (3,228)
These higher-order
= 2.33, p
Intentionality
Immediate x Importance
interaction of Immediate
short of statistical
< .10.
interactions
confirm the visual impression from Figure 1.
The delay group outperforms the immediate group only in the intentional condition.
9 Separate analysis of variance on the intentional subjects did result in a significant main effect for Immediate-Delay,
F (1,38) = 7.39, R < .001 and
the Importance Level x Immediate Delay interaction was also reliable, F (3,114) = 8.16, p < .001.
The improvement in the delay group was entirely on the two
highest levels of importance. was not reliable.
The slight improvement at the lower two levels
In the incidental condition the effect of immediate-delay
was not significant; performance between the immediate and delay groups was comparable. The data suggest that students in
the intentional study condition were
able to use the extra time provided to enhance their recall, but the improvement shown by those awarded extra time was not uniformly distributed across importance level.
Students used extra study time to improve their recall of
the important elements of texts, thus producing the anticipated diagnosis recall pattern.
That intentional study strategies are involved in
ment is supported by the incidental-intentional comparison.
this improve-
Although both
delayed recall groups interacted with the story for the same amount of time, and the incidental orienting task would be regarded as semantic (Craik & Lockhart,
1972),
only the deliberate study strategies of the intentional group
led to enhanced recall of important units.
In addition, 80% of the delayed
intentional group reported using some recognizable strategy to help their recall.
Only two of the 20 incidental delayed group reported awareness that
a recall would probably be called for and only one subject indicated a surreptitious plan for remembering.
Unfortunately we did not retain the
physical records of the students studying, e.g., sections of text. obvious later.
notes or underlined
This oversight we have reason to regret, as will become
10 In Experiment la the comparison between immediate and delay conditions was a between subjects variable.
This was done to ensure comparability with
the incidental groups where a between subjects manipulation was of course necessary to maintain credibility of the cover story,
In Experiment lb we
repeated the intentional condition with further groups of college suudents as a desirable replication, own recall if
and to see whether individual students improve their
given extra time to process the material. Experiment lb
Method Subjects. in this study.
A further group of 40 college student volunteers participated Half the students were female.
Stimulus materials. Procedure. individually.
These were the same as in Experiment la.
Each student
was tested on two separate days, in groups or
Half the students were randomly assigned to the Cat story on
Day 1, and the Dragon story on Day 2, and the reverse was true for the On the first day, they listened to the story while
remaining subjects.
simultaneously reading it
through and then, after a short retention interval
(5 minutes, during which they worked on a word puzzle), they attempted gist recall.
Following this they were given five minutes extra study and told to
undertake any activity they wished in order to improve their recall.
They
had at their disposal note pads, felt pens, pens and a copy of the text printed in primary type.
After the five-minute period had elapsed the aids
were removed and the students attempted gist recall, again following a fiveminute filled retention interval.
On the second day the entire procedure
was repeated with the second story, but before the study period the students were told that it helps some people to underline or take notes and they might do so if they wished.
The protocols were coded and scored blind by two
11 independent raters for gist recall of idea units (interrater reliability = .94). Results and Discussion Again there were no obvious effects of story or sex of subject so the data were combined over these factors.
A 2 (Immediate-Delay) x 2 (Prompt, No Prompt)
x 4 (Importance Level) mixed analysis of variance revealed a main effect of Immediate-Delay, F (1,38) = 68.35, p < .001 and of Importance Level, F (3,114) = 295.,
a
-s